Cement and Concrete Research 41 (2011) 609–614

Contents lists available at ScienceDirect

Cement and Concrete Research

j ourna l homepage: ht tp: / /ees.e lsev ie r.com/CEMCON/defau l t .asp

Comparison between surface and bulk hydrophobic treatment against corrosion ofgalvanized reinforcing steel in concrete

F. Tittarelli ⁎, G. MoriconiDepartment of Materials and Environment Engineering and Physic, Polytechnic University of Marche, 60131 Ancona, Italy

⁎ Corresponding author. Tel.: +39 071 2204732; fax:E-mail address: f.tittarelli@univpm.it (F. Tittarelli).

0008-8846/$ – see front matter © 2011 Elsevier Ltd. Aldoi:10.1016/j.cemconres.2011.03.011

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 July 2010Accepted 21 March 2011

Keywords:Concrete (E)Corrosion (C)Reinforcement (D)Zinc (D)Hydrophobic treatment

The effectiveness of bulk hydrophobic treatment against corrosion of galvanized steel reinforcement inconcrete specimens with w/c=0.45 and w/c=0.75 was compared with that of surface treatment, even in thepresence of cracks 0.5 and 1 mmwide in the concrete cover. In this case surface hydrophobic treatments wereapplied both before and after cracking as a preventive and a restorative method against reinforced concretedeterioration, respectively. The obtained results in terms of water absorption, electrochemical measurements,chlorides penetration, and visual observations carried out on reinforced concrete specimens during theexposure to wet–dry cycles in 10% NaCl solution showed that bulk hydrophobization is the most effectivetreatment in improving the corrosion resistance of galvanized steel reinforcements in concrete also in thepresence of cracks. Surface hydrophobization is very effective just in the first few exposure cycles to theaggressive environment and when used as a restorative method which is able to cancel the deleterious effectof cracks only 0.5 mm wide.

+39 071 2204729.

l rights reserved.

© 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Weathering is the most important harmful factor for buildingmaterials causing their deterioration, detectable through exfoliation,scaling, crumbling and reinforcement corrosion. Weathering is ingeneral related to water penetration into porosity of materials, sincewater acts as the main carrier of aggressive agents, as chlorides forreinforcement corrosion, and water is essential for most deteriorationreactions [1].

A water droplet in contact with concrete, that is a poroushydrophilic material, wets the solid surface by spreading itself andpenetrates the cementitious matrix by means of capillary forcesfollowing the Washburn equation, p=(2 γ/rc) cos σ, where γ is theliquid surface tension, rc is the capillary pore radius, and σ is thecontact angle. The contact angle is less than 90°, p is positive andwater fills the pore spontaneously. Since the molecular attractionbetween water and the concrete pore walls may be lowered by usinghydrophobic agents, such as those currently named silanes andsiloxanes [2], surface hydrophobic treatments or the introduction ofhydrophobic agents directly in the mixture should involve an increasein durability. Previous works have shown that hydrophobic admix-tures protect steel reinforcement from corrosion only in soundconcrete, while in cracked concrete it can induce a catastrophiccorrosion [2]. This unexpected result has been ascribed to a greater

oxygen diffusion through an unsaturated hydrophobic cementitiousmatrix that, in this way, can feed more quickly the cathodic reactioncontrolling the corrosion process [3]. However, if the reinforcement isgalvanized, in which case it has been already experimented thatpassivation is mainly promoted by oxygen [4], the use of hydrophobicadmixture always improves the corrosion resistance of galvanizedsteel reinforcements in concrete, even in the presence of concretecracks, especially when high w/c ratios are used [5]. In this work, theeffectiveness of bulk hydrophobic treatment against corrosion ofgalvanized steel reinforced in concrete specimens with w/c=0.45andw/c=0.75was comparedwith that of the surface treatment, evenin the presence of cracks 0.5 and 1 mm wide. In this case surfacehydrophobic treatments were applied both before and after crackingas a preventive and a restorative method against reinforced concretedeterioration, respectively.

2. Experimental

2.1. Materials

A commercial Portland-limestone blended cement type CEM II/A-L32.5 R according to EN-197/1 was used. Natural sand with 6 mmmaximum size and gravel with 11 mm maximum size were used asfine and coarse aggregates, respectively, to manufacture concretes. A45% aqueous emulsion of an alkyl-triethoxy-silane was used as ahydrophobic silane-based admixture to manufacture bulk hydropho-bic cementitious materials. A commercial hydrophobizing agent

Table 1Concrete mixture proportions (kg/m3) and related compressive strengths (MPa).

Mixtures w/c=0.45 w/c=0.45hydrophobic

w/c=0.75 w/c=0.75hydrophobic

Water 240 233 240 236Cement 533 533 320 320Silane – 11.8 – 7.1Sand 639 639 949 949Gravel 847 847 716 716Rc at 2 days 19.9 16.8 6.7 4.,9Rc at 7 days 35.4 27.5 14.5 9.2Rc at 28 days 39.2 32.3 25.6 19.8

610 F. Tittarelli, G. Moriconi / Cement and Concrete Research 41 (2011) 609–614

based on siloxanic resin was used as a hydrophobic surface treatmenton concrete specimens.

2.2. Specimens

Concrete mixtures with w/c of 0.45 and 0.75 were manufacturedwith and without silane as a hydrophobic admixture added at adosage of 1% of active ingredient by cement weight. The proportionsof the concrete mixtures are given in Table 1, where compressivestrengths at 2, 7, and 28 days are also reported. In the presence ofsilane admixture, a reduction in the concrete compressive strength ofabout 18 and 23% was recorded with respect to that of non-hydrophobic concrete when a w/c ratio of 0.45 and 0.75 was used,respectively.

For each concrete type, prismatic specimens (70×70×280 mm insize) were produced (Fig. 1). These prismatic specimens werereinforced with a hot dip galvanized steel plate (210×40×1 mm)embedded at mid depth from a specimen side. The zinc coating,obtained by immersion in molten zinc, was 100 μm thick with anouter pure zinc layer about 20 μm thick. An electric cable, isolatedthrough a PVC sheath, was connected to each steel plate by means ofspot welding. The weld was protected by epoxy resin before castingthe concrete mixture. Steel plates instead of usual rebars were used inorder to facilitate the cracking of specimens and tomodulate the crackwidth.

After casting, all the specimens were wet cured for 2 days(RH=100% and T=22 °C) and air dried for 1 month at RH=60%and T=22 °C. Then, some specimens for each concrete type weredeliberately cracked by a flexural stress in order to open a crack 0.5 or1 mmwide in the middle zone, with the crack apex reaching the steelreinforcement. In order to localize the crack formation at apredetermined point, a preformed V-notch, 10 mm deep and 14 mmwide, was produced on these specimens by placing a plastic rod on thebottom surface of the mold. Once the crack of required width,measured by an extensometer, was obtained its opening wasmaintained after the elimination of the flexure load by the insertionof small elements of hard polymeric material of suitable thickness. Inorder to make the oxygen and chloride flow unidirectionally andperpendicularly directed to the steel plate through the concrete cover,

Fig. 1. Prismatic reinforced

all the other sides of the specimens but the one containing the crackapex were epoxy coated.

The hydrophobic surface treatment, if any, was applied after curingof the concrete specimens by impregnating the specimens' surfaceuntil saturation with the commercial hydrophobizing agent. Somespecimens were cracked by flexural stress (crack width of 1 mm and0.5 mm) after an additional week. In order to estimate the efficiencyof the hydrophobic surface treatment as a restorative method againstdeterioration, half of the specimens were coated after cracking. Inorder to estimate the efficiency of the surface coating as a preventivemethod against corrosion, the other half was coated before cracking.Table 2 explains the nomenclature used to identify the different casesstudied.

2.3. Tests

2.3.1. Water absorption at atmospheric pressureWater absorption tests at atmospheric pressurewere carried out on

concrete andmortar specimens according to the procedure reported inUNI 7699:2005— Testing hardened concrete— determination ofwaterabsorption at atmospheric pressure.

2.3.2. Electrochemical measurementsAfter the curing period, the hydrophobic surface treatment and the

cracking (about 45 days from the cast), all the specimens wereexposed to wet–dry cycles (2 days wet followed by 5 days dry) in a10% NaCl aqueous solution for four months.

The corrosion risk of the reinforced concrete specimens exposed tothe chloride environment was evaluated by free corrosion potentialmeasurements with respect to a saturated calomel electrode (SCE) asreference, while the kinetics of the corrosion process were followed bypolarization measurements. The polarization resistance, in inverserelation with the corrosion rate, was measured through the galvano-dynamic method (0.5 μA/s) by calculating the average value betweenthe anodic and the cathodic branch (ΔV=±5mV). An externalgraphite bar (210×40×10 mm) on the surface of the specimen wasused as counter-electrode; the surface area of galvanized steel exposedto the environment was 210×40×2 mm2. The electrochemicalmeasurements reported in the following graphs were carried outevery cycle at the end of the wetting period with the specimen, thereference electrode and the counter-electrode immersed in the chloridesolution.

2.3.3. Chloride content and visual evaluationIn all cases, the free chloride content of the concrete in proximity of

the steel reinforcementwas evaluated bywater extraction at the end ofthe exposure time to the chloride environment. The concrete powderwas collected by a drill, at 0–1 cm from the galvanized plate and 1 cmfrom the notch initiating crack, fromboth the two sides of the crack. Thepowderwasmixedwithdistilledwater (water/powder=12byweight)for 24 hours, then the solution was filtered and analyzed by ionchromatography. The chloride content has been reported as % by

concrete specimens.

0

1

2

3

4

5

6

7

8

9

0 20 40 60 80 100 120 140 160 180

Time (hours)

Abs

orbe

d w

ater

(%

)

no-h no-h-0.5 no-h-1h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur

Fig. 3. Water absorbed by concrete specimens with w/c=0.75, as a function of the testtime.

Table 2Nomenclature used to identify the different cases studied.

Description Nomenclature

No hydrophobic treatment, no crack no-hNo hydrophobic treatment, crack 0.5 mm wide no-h-0.5No hydrophobic treatment, crack 1 mm wide no-h-1Bulk hydrophobic treatment, no crack h-bulkBulk hydrophobic treatment, crack 0.5 mm wide h-bulk-0.5Bulk hydrophobic treatment, crack 1 mm wide h-bulk-1Surface hydrophobic treatment, no crack h-surSurface hydrophobic treatment applied before crack 0.5 mm wide h-sur-0.5Surface hydrophobic treatment applied before crack 1 mm wide h-sur-1Surface hydrophobic treatment applied after crack 0.5 mm wide h-0.5-surSurface hydrophobic treatment after crack 1 mm wide h-1-sur

611F. Tittarelli, G. Moriconi / Cement and Concrete Research 41 (2011) 609–614

cement weight. The cement content of themanufactured concreteswascalculated by the mix proportions reported in Table 1.

Finally, in order to validate and to complete the evaluation of theelectrochemical behavior, the galvanized steel plates were extractedby splitting the concrete specimens after 16 wet–dry cycles in thechloride solution and the corrosion extent was assessed by visualobservation. Zinc corrosion products were identified by X-raydiffraction.

3. Results and discussion

3.1. Water absorption at atmospheric pressure

Figs. 2 and 3 show the results of water absorption tests atatmospheric pressure obtained on concrete specimens with a w/cratio of 0.45 and 0.75, respectively. It is evident that cracks in theconcrete cover do not generally affect the test results of non-hydrophobic concretes and bulk hydrophobic concretes, probablybecause the surface area of the crack is much smaller with respect tothe whole specimen surface in contact with water. At the end of thetest, hydrophobic concretes with silane admixture at a dosage of 1% bycement weight absorb just 50% and 35% of water absorbed by thecorresponding non-hydrophobic ones when w/c of 0.45 and 0.75 areused, respectively. The hydrophobic admixture due to its ability todecrease the molecular attraction between water and the concreteporewalls, without decreasing the concrete porosity [2], seems able tocancel the deleterious effect of a high w/c and concrete cover crackingup to 1 mm wide on concrete water absorption.

On the other hand, surface hydrophobic treatment seems to beeffective just in the first days of full immersion in water but afterlonger water contacts surface hydrophobic concrete absorbs almostthe same water absorbed by the non-hydrophobic one especiallywhen high w/c ratios are adopted and the surface treatment is used asa preventive method against deterioration (before cracking). More-

0

1

2

3

4

5

6

0 20 40 60 80 100 120 140 160 180

Time (hours)

Abs

orbe

d w

ater

(%

)

no-h no-h-0.5 no-h-1h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur

Fig. 2. Water absorbed by concrete specimens with w/c=0.45, as a function of the testtime.

over, when the hydrophobic surface treatment is used as a restorativemethod against deterioration (after cracking) it is able to hinderwaterpenetration only in sound concrete or in a 0.5 mm cracked concretebut not in a 1 mm cracked concrete.

3.2. Electrochemical measurements

The discussed electrochemical values are averaged among themeasurements carried out on three specimens of each concrete type.

Just after the cast, the reinforcements embedded in the specimensassumed potentials values of about −1350 mV/SCE which are typicalvalues for the zinc activation state from the cast time as long as thespecimens were stored in their molds. After demolding the valuesrose after about 6 hours to −750 mV(SCE) in concretes with w/c=0.45 and to−600 mV/SCE in concretes with w/c=0.75. Althoughthese potentials would represent active corrosion for steel they arenot for zinc. This difference related to w/c is due to the greaterporosity of the cementitious matrix with the higher w/c. The greaterporosity allows a higher amount of oxygen to flow through the concretecover, to reach the galvanized rebars and to induce a better zincpassivation. During the whole air curing period the potentials continueto become more positive (from −700 mV/SCE to −500/−400 mV/SCE). This behavior was observed regardless of the presence of thehydrophobic treatment and it is analogous to those already discussed inReferences [4,6] and [7].

Fig. 4 shows the free corrosion potential values of galvanized steelplates embedded in concrete specimenswithw/c=0.75 as a function ofwet-dry cycles in the aggressive environment based on chlorides. Afterexposure to chlorides, in sound non-hydrophobic concrete thepotentials assume active values of about −1000 mV/SCE reflecting a

-1200

-1100

-1000

-900

-800

-700

-600

-500

-400

1 2 3 4 5 6 7 8 9 10 11 12

Cycles

E (

mv/

SC

E)

h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur no-hno-h-0.5 no-h-1

Fig. 4. Corrosion potential of the galvanized steel plates embedded in concretespecimens with w/c=0.75, as a function of the test time.

0

5

10

15

20

25

30

1 2 3 4 5 6 7 8 9 10 11 12

Cycles

Rp

(Ohm

m2 )

h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur no-hno-h-0.5 no-h-1

Fig. 7. Polarization resistance of the galvanized steel plates embedded in concretespecimens with w/c=0.45, as a function of the test time.

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

h-bulk h-bulk-0.5 h-bulk-1 h-sur h-sur-0.5 h-sur-1 h-0.5-sur h-1-sur no-h no-h-0.5 no-h-1

Chl

orid

es (

% b

y ce

men

t wei

ght)

a

0

5

10

15

20

25

30

1 2 3 4 5 6 7 8 9 10 11 12

Cycles

Rp

(Ohm

m2 )

h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur no-hno-h-0.5 no-h-1

Fig. 5. Polarization resistance of the galvanized steel plates embedded in concretespecimens with w/c=0.75, as a function of the test time.

612 F. Tittarelli, G. Moriconi / Cement and Concrete Research 41 (2011) 609–614

generally great corrosion risk and they remain constant for almost theentire test period. Galvanized steel in the non-treated concrete does notpassivate due to the lack of oxygen, that is the fundamental oxidizingagent for a fast and effective zinc passivation in concrete [3–5], in a wellsaturated cementitiousmatrix as thatwith ahighw/c (0.75) after 2 daysof immersion in a chloride solution. In the presence of concrete cracks,but without the hydrophobic treatment, the free corrosion potentialafter 7 cycles moves towards more anodic values (−800 mV/SCE, thatrepresents active corrosion for steel) when, due to the chloride attack,the underlying steel becomes exposed at the intersection with thecracks and then it actively corrodes.

On the other hand, in hydrophobic concrete, whatever the type ofhydrophobic treatment, just after the exposure to the aggressiveenvironment the galvanized steel plates assume values of −600 mV/SCE in sound specimens or in after cracking surface treated specimensif the crack width is 0.5 mm, and−800 mV/SCE in all the other cases.Thismeans the ability of the hydrophobic treatments to prevent waterpenetration especially in sound concrete or when used as a restorativemethod in 0.5 mm crack width. However, when bulk hydrophobictreatment is used, the free corrosion potentials move towards morepassive values, indicating a very low corrosion risk, after a few wet-dry cycles. This is due to the great capacity of the hydrophobicadmixture to repel water, even with the presence of concrete cracks,which favors faster oxygen diffusion through concrete, thus makinggalvanized reinforcement passivation faster and easier [3–5]. On thecontrary, when surface hydrophobic treatment is applied as apreventive or as a restorative method against corrosion, the corrosionpotential moves towardsmore active values due to the chloride attackto zinc and at the end of the test it reaches the same values detected innon-hydrophobic concrete. This is due to the lower capacity of the

-1200

-1100

-1000

-900

-800

-700

-600

-500

-400

1 2 3 4 5 6 7 8 9 10 11 12

Cycles

E (

mv/

SC

E)

h-bulk h-bulk-0.5 h-bulk-1h-sur h-sur-0.5 h-sur-1h-0.5-sur h-1-sur no-hno-h-0.5 no-h-1

Fig. 6. Corrosion potential of the galvanized steel plates embedded in concretespecimens with w/c=0.45, as a function of the test time.

surface treatment to hinder water penetration for a long time alsowhen applied as a restorative method against deterioration and in theabsence of concrete cracks.

Concerning the kinetics of the corrosion process, in the absence ofconcrete cracks, the polarization resistance (Fig. 5) of the steelreinforcement in bulk hydrophobic concrete with w/c=0.75 can befour times higher with respect to that in non-hydrophobic concrete,indicating the good efficiency of the hydrophobic admixture indecreasing the corrosion rate; moreover, this value seems to furtherincrease during the test time indicating that zinc passivation increasesprogressively. In the presence of concrete cracks, the differencebetween bulk hydrophobic and non-hydrophobic concrete is moder-ate, but the polarization resistance of steel reinforcement inhydrophobic concrete remains at least three times higher with

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

h-bulk h-bulk-0.5 h-bulk-1 h-sur h-sur-0.5 h-sur-1 h-0.5-sur h-1-sur no-h no-h-0.5 no-h-1

Chl

orid

es (

% b

y ce

men

t wei

ght)

b

Fig. 8. a. Free chloride ion concentration (% by cement mass) in concrete specimenswith w/c=0.45. b. Free chloride ion concentration (% by cement mass) in concretespecimens with w/c=0.75.

613F. Tittarelli, G. Moriconi / Cement and Concrete Research 41 (2011) 609–614

respect to that of non-hydrophobic concrete. Moreover, in this case,cracked hydrophobic concrete behaves even better than sound non-hydrophobic concrete, whatever the crack width; therefore, bulkhydrophobic treatment seems to cancel, from a corrosion point ofview, the detrimental effect of concrete cover cracking.

The two different crack widths do not affect the corrosion resultsin non-hydrophobic concrete specimens with w/c=0.75, due to thehigh porosity of the cement matrix, while, in the hydrophobic ones, aslightly higher corrosion rate is detected in the presence of the widercrack. When hydrophobic surface treatment is adopted, initially abetter corrosion behavior is detected just in the absence of concretecracks or in the presence of 0.5 mm wide cracks covered by thesurface treatment. However, after 7 cycles surface hydrophobicconcrete behaves almost like non-hydrophobic concrete whateverthe surface treatment.

A good quality concrete matrix with a w/c as low as 0.45 seems topartially hide the beneficial effect of the hydrophobic treatment. Thecorrosion risk described by the free corrosion potential measurements(Fig. 6) is about the same regardless of the cement matrix, especiallyat the end of the test. However, the polarization resistance of steelreinforcement (Fig. 7) is always higher in the bulk hydrophobicconcrete specimens with respect to that of the corresponding non-hydrophobic ones, even if the difference is not as marked as observedin the more porous cement matrix. In particular, in the absence ofconcrete cracks, bulk hydrophobic concrete shows a polarizationresistance that is approximately twice that of other specimens. Whenconcrete cover is cracked the difference is not so evident. However,only bulk hydrophobization and surface hydrophobization applied asa restorative method seem to give a certain increase in polarizationresistance. Moreover, due to the low porosity of the cement matrix,the crack width influenced the test results both in the presence and inthe absence of silane with corrosion risks and corrosion rates whichgenerally increase with the crack width.

Finally, by comparing the results obtained in concrete specimenswith w/c=0.75 (Figs. 4 and 5) and w/c=0.45, (Figs. 6 and 7) it isevident that the hydrophobic admixture, even in the presence ofconcrete cracks, decreases the corrosion rate monitored in veryporous concrete, manufactured with w/c=0.75, to values comparablewith those obtained in good quality concrete, manufactured with w/c=0.45. In other words, silane cancels the detrimental effect, at leastfrom the corrosion point of view, of a large porosity of the cementmatrix.

3.3. Chloride content and visual evaluation

The chloride content values reported in the graphs are averagedamong the values obtained by the analysis carried out on threespecimens of each concrete type.

In concretes with w/c=0.45 (Fig. 8a) the chloride values rangedfrom 0.2% by cement weight in sound hydrophobic concretes(regardless the type of treatment) to 3% by cement weight in non-hydrophobic concretes or cracked concretes with the hydrophobicsurface treatment. In concretes with w/c=0.75 (Fig. 8b), the chloridevalues generally ranged from 0.7% by cement weight in soundhydrophobic concrete (regardless the type of treatment) to 9.0% bycement weight in non-hydrophobic concretes or cracked concreteswith the hydrophobic surface treatment. In particular, in the presenceof concrete cracks only bulk hydrophobic treatment is always able toreduce chloride penetration (Fig. 8a and b) significantly by at least

Fig. 9. a. Visual observation of the galvanized steel plate embedded in concrete specimenswith w/c=0.75 in the absence of the hydrophobic admixture (up: no crack; middle:0.5 mm crack; down: 1 mm crack). b. Visual observation of the galvanized steel plateembedded in concrete specimenswith w/c=0.75 in before cracking surface hydrophobicconcrete (up: no crack;middle: 0.5 mmcrack; down: 1 mmcrack). c. Visual observation ofthe galvanized steel plate embedded in concrete specimens with w/c=0.75 in aftercracking surface hydrophobic concrete (up: 0.5 mm crack; down: 1 mm crack). d. Visualobservation of the galvanized steel plate embedded in concrete specimenswithw/c=0.75in thepresenceof thehydrophobic admixture (up: no crack;middle: 0.5 mmcrack; down:1 mm crack).

Fig. 10. X-ray Diffraction analysis of white zinc corrosion products.

614 F. Tittarelli, G. Moriconi / Cement and Concrete Research 41 (2011) 609–614

about 50% and 70% when w/c=0.45 (Fig. 8a) and w/c=0.75 (Fig. 8b)are used, respectively.

However, when concrete with high w/c ratio is used (w/c=0.75)also the surface treatment seems effective in reducing chloridepenetration when used as a restorative method in 0.5 mm concretecover crack width (Fig. 8b).

The corrosion attack evaluated by the visual observations carriedout on the galvanized steel plates removed from the specimens afterthe exposure to the chloride environment were in quite goodagreement with the conclusions obtained by the water absorptiontest (Section 3.1), the electrochemical measurement (Section 3.2) andthe chloride content analysis.

The galvanized steel plates extracted from non-hydrophobicconcrete (Fig. 9a) showed a very deep general corrosive attack withred rust appearing on the surface indicating that total consumption ofthe protective zinc layer had occurred.

Concerningwith the hydrophobic surface treatment, when appliedas a restorative method (Fig. 9b and c), dark Fe–Zn alloys appeared onthe plate, with zinc grains no longer visible on the surface, meaningthe total consumption of the pure zinc layer, both in sound and in0.5 mm wide cracked concrete. In the other cases, light red rustappeared also on the surface probably due to the corrosion of Fe fromalloying sublayers since Fe–Zn alloys are not stable in presence ofchlorides [7,12].

On the other hand, zinc grains were still visible in steelreinforcement embedded in bulk hydrophobic concrete (Fig. 9c) andclose to the crack apex a white surface deposit of adherent andcompact zinc corrosion products, later identified by X-ray diffractionas zinc oxides, hydroxides and calcium hydroxyzincate (Fig. 10),appeared. In particular calcium hydroxyzincate is a well knownpassivating zinc corrosion product that, once formed, protects theunderlying pure zinc layer from further corrosion [4–11].

For brevity, only photos of the galvanized steel plates embedded inconcretes with w/c=0.75 are reported. However a very similar trend

was detected from the visual evaluation of the galvanized steelreinforcements removed from the specimens with w/c=0.45.

4. Conclusions

The effectiveness of bulk hydrophobic treatment against chlorideinduced corrosion of galvanized steel reinforcement in concretespecimens with w/c=0.45 and w/c=0.75 was compared with that ofthe surface hydrophobic treatment, even in the presence of cracks 0.5and 1 mmwide. The surface hydrophobic treatment was applied bothbefore and after cracking as a preventive and a restorative methodagainst reinforced concrete deterioration, respectively. The obtainedresults showed that bulk hydrophobic treatment is the most effectivetreatment to improve the corrosion resistance of galvanized steelreinforcements in concrete, also in the presence of cracks. Especiallywhen high w/c are adopted, surface hydrophobic treatment iseffective in preventing water penetration just for short period (fewdays) of contact with water. Moreover, the hydrophobic surfacetreatment is able to improve significantly the corrosion behavior ofgalvanized reinforcements in concrete only in the first few exposurecycles and when used as a restorative method in the presence of0.5 mm wide cracks.

References

[1] C.L. Page, K.W.J. Treadaway, Aspects of the electrochemistry of steel in concrete,Nature 297 (1982) 109–115.

[2] F. Tittarelli, G. Moriconi, The effect of silane-based hydrophobic admixture oncorrosion of reinforcing steel in concrete, Cement and Concrete Research 38 (11)(2008) 1354–1357.

[3] F. Tittarelli, Oxygen diffusion through hydrophobic cement-based materials,Cement and Concrete Research 39 (10) (2009) 924–928.

[4] F. Tittarelli, T. Bellezze, Investigation of the major reduction reaction occurringduring the passivation of galvanized steel rebars, Corrosion Science 52 (3) (2010)978–983.

[5] F. Tittarelli, G. Moriconi, The effect of silane-based hydrophobic admixture oncorrosion of galvanized reinforcing steel in concrete, Corrosion Science 52 (9)(2010) 2958–2963.

[6] T. Bellezze, M. Malavolta, A. Quaranta, N. Rufini, G. Roventi, Corrosion behaviour inconcrete of three differently galvanized steel bars, Cement and Concrete Composites28 (2006) 246–255.

[7] Z.Q. Tan, C.M. Hansson, Effect of surface condition on the initial corrosion ofgalvanized reinforcing steel embedded in concrete, Corrosion Science 50 (9) (2008)2512–2522.

[8] S.R. Yeomans, Performance of black, galvanized, and epoxy-coated reinforcingsteels in chloride-contaminated concrete, Corrosion Engineering (January 1994)72–81.

[9] G. Arliguie, J.P. Ollivier, J. Grandet, Study on the retarding effect of zinc onhydration of Portland cement paste, Cement and Concrete Research 12 (1) (1982)79–86.

[10] A. Macias, C. Andrade, Corrosion of galvanized steel in diluted Ca(OH)2 solutions(pH 11), British Corrosion Journal 22 (3) (1987) 162–171.

[11] A. Macias, C. Andrade, The behavior of galvanized steel in chloride containingalkaline solution. I. The influence of cation, Corrosion Science 30 (4/5) (1990)393–407.

[12] E. Maahn, B. Soresen, The influence of microstructure on the corrosion propertiesof Hot-Dip galvanized reinforcement in concrete, Corrosion 42 (4) (1986) 187.